Shimano 12 speed road bike shifter conversion

These are parts to upgrade a Shimano 105 ST-5800 road bike STI shifter from 11 to 12 speed
8
18
0
917
updated May 4, 2023

Description

PDF

Last updated 20 January 2023

What this project is

These are parts to upgrade a Shimano 105 ST-5800 road bike STI shifter from 11 speed to 12 speed. They may also work on other Shimano 11-speed shifters of the same generation, but the next generation requires the remixed part from jr_bike.

Parts Needed

Only two pieces are needed: one PR part (for Primary Ratchet) and one SR part (for Secondary Ratchet). It would be nice if the SR part was the same series as the PR, but the original Secondary Ratchet will probably work too.

It's highly recommended to use a high quality cable and have newer housings. Shorter cable pull means the shifting is more sensitive to cable stiction that higher quality, die drawn cables try to reduce.

My nomenclature has changed. Instead of referring to parts in terms of their size to an arbitrary other part, now they will be by the angle between teeth. 11 speed was about 16.1 degrees. 12 speed is between 15.4 and 15.6 degrees.15.4 degrees will work better for a Shimano cog-to-cog pitch of about 3.55mm. 15.6 degrees should work better for a cassette more towards a SRAM cog pitch of 3.65mm. Or just split the difference at 15.5 degrees and use the PR 15.5deg 20.2dia 10pct R1.dxf and SR 15.5deg v3r1.dxf parts.

I'm not completely sure that the 20.2mm diameter is correct. 20.4mm was a bit too big and I needed to file down the backsides of the teeth. The balancing act of tooth spacing angle and tooth tip to tip height is briefly covered in the naming convention section.

Installation

Conversion of the shifter is not for the faint of heart. The shifter needs to be almost completely disassembled in order to replace the gear ratchets. I have a YouTube video of the disassembly, a separate video covering reassembly, and a brief run through the gears with my 15.75 degree version. The most challenging steps are attaching pawl springs as parts are pushed together and aligning pins to holes while those pawl springs are trying to push them out of alignment. In addition to changing out the the parts, the edge of the aluminum cable spool needs to be notched or filed down so it can rotate farther before hitting an endstop and pull more cable.

Material

Due to the impact of the pawl hitting the gear teeth while upshifting, the Primary Ratchet part should be made of a strong material like metal. Though the Secondary Ratchet doesn't have to withstand as much impact, the gear teeth still experience high loads and should probably also be a durable material like metal, if only so it doesn't need replacing months later.

Fabrication

I sent my parts out to be laser cut, so I have included DXF part files which most laser cutting services will accept for parts. My concern about using laser cutting for metal is the edges are rough and might wear away at the pawls that rub against them. To mitigate this, the faces of the gear teeth can be smoothed with light use of a miniature file or emery disc.

If you have access to one, a water jet cutter would probably leave a better cut edge, and wire EDM would just be overkill.

The PR parts are nominally 2.9mm thick and the SR are 2.5mm. If not using the spacer washer, the SR part can be 2.8mm thick, so in the interest of saving money, I got both og my latest parts cut out of 3mm wear resistant steel and filed them thinner. As it is, the PR part also needs a flat part cut out of the back too. Check the STL model for the location of the feature. It doesn't have to be that exact- it can easily be done with a rotary tool with an emery cutting disc, fitting it to the cable spool for fit.

Information is encoded in the bumps on the edges of the ratchets. For the primary ratchet, they can be left as is. On the secondary ratchet, they jut out too much, will probably interfere with the pawls, and should be filed down. I made them that big to reach the minimum size limit of my laser cutting service's automated quoting tool.

I apologize for providing so many different parts. Though I have found something that works adequately for me, I have made other adjustments that should work better, but I haven't yet tested, which is why this project is labeled as a work in progress. Also, cassettes may vary in their spacing, so these allow for some variation. I used a third-party 12-speed cassette, Shimano CN-M7100 SLX 12-speed chain, and Shimano RD-5800-SS 105 short cage 11-speed derailleur that can barely handle the 32 tooth large cog. When Shimano 12-speed road cassettes become available, the cog pitch may slightly differ from other cassettes.

About my updated naming convention

PR parts. Short for Primary Ratchet, this is the gear that holds the cable spool in position and is responsible for the amount of movement per gear.

  • The first number (usually 15.4, 15.5 or 15.6deg) is the angle between the middle teeth. This is what really determines the amount of cable pull. The angle between the first and second tooth (for the smaller cogs) is larger to give a bit of extra travel for the first click. This really should not matter since the derailleur limit screw will prevent using the whole range of travel.
  • The second number is the diameter from tooth tip to tooth tip. On the 11-speed gears, this is about 20.8mm, but because 12-speed spaces the teeth closer, there is the risk of a pawl hitting the top of a tooth instead of sliding right behind it. Decreasing the diameter helps prevent this problem. This problem should really be fixed by decreasing the pawl length, but that is an extra, more complicated part to make, so we kludge this workaround instead. The problem is if the diameter is too small, pressing slowly to make an upshift might cause the ratchet to skip teeth. I have not determined the best diameter for any given tooth angle.
  • The third number is the percentage extra spacing to the largest cog. This has arbitrarily been chosen as 10%. Much like with the smallest cog, a little extra cable pull is added insurance to make sure that cog is reached in case the cable pull was minutely too little over the previous ten cogs. Again, the limit screw should be set to prevent the chain from going too far. Any extra cable pull will get absorbed in the system by stressing parts a little like stretching the cable or compressing the housing.

SR Parts. What I refer to as the Secondary Ratchet, this part is what a pawl catches on to push the primary ratchet to a larger cog. The spacing on it is not as important. The SR part name just has the gear tooth angle. Matching the gear tooth angle between the PR and SR parts is just so that the downshift lever always catches at the same point, but it's not like many people will be able to tell the difference of tenths of a degree, so any SR part could be used. It is probably even possible to use the original 11-speed secondary ratchet to be economical, but the lever will need to be pushed a bit farther for the 11-12 largest cog shift. Rationalizing it, this would be a good indication that the last cog has been reached.

Model File Information

I have included a Fusion 360 file for both the parts. Most of the parameters are defined and changeable without digging into individual sketches. I have even starred the parameters that likely will be changed most often and corresponding to the above naming convention:

*The angle between teeth- "DegreesPerClick." This used to be a messy formula based on a lot of irrelevant factors and assumptions but is now just an angle that should be between 15.4 and 15.8 degrees. Work outside that range at your own risk.

  • The tip-to-tip diameter of the Primary Ratchet (labeled as "PRTipToTipDia"). Going larger than 20.64 may cause some unanticipated errors. if you truly want to increase it past that, make PR_MajorDia larger too, but this will be to large for the tooth angle less than 15.7 degrees.
  • The amount of extra pull for the last cog, named "LastClickOversize" 10% might be a bit too much stretch for the cable, so it could be decreased to 0.05 (5%) or even set to zero. This might also make it feel just slightly different in an upshift from the third largest cog to the next smallest, if you find that annoying, or it results in some unanticipated behavior, the number can be decreased. Note that a negative number shouldn't really do anything except maybe cause a rebuild error.

Information Encoding

The ratchets have information encoded in the bumps on them to make it easier to distinguish between parts if getting many made at the same time. My laser cutting service had a minimum order, so I had parts cut with different angles and diameters.

The numbers on the bumps are in base 3. A smooth bump, divot, or lack of a feature represents a zero, a triangle (one edge on top) is a one, and a rectangle or trapezoid (two edges on top) is a two. Reading clockwise from the flat on the primary ratchet is first the decimal portion of the tooth angle (15.4, 15.5, or 15.6), then comes the decimal portion of the diameter (20.0, 20.2, 20.4). After that is a gap and then the percentage extra spacing in multiples of five which so far is only 5x2%, and finally, the revision1. The secondary ratchet has only the revision, and the tooth angle.

Tags



Model origin

The author marked this model as their own original creation. Imported from Thingiverse.

License